label-studio¶

Description¶

Label Studio is an open source data labeling tool that supports multiple projects, users, and data types in one platform. It allows you to do the following: Perform different types of labeling with many data formats. As well as integrate Label Studio with machine learning models to supply predictions for labels (pre-labels), or perform continuous active learning.

Environment Modules¶

Run module spider label-studio to find out what environment modules are available for this application.

Environment Variables¶

HPC_LABELSTUDIO_DIR - installation directory
HPC_LABELSTUDIO_BIN - executable directory

Additional Usage Information¶

Warning

Label Studio will use /home/<username>/.local/share/label-studio as the default data directory if not specified. The home directory on HiPerGator has a 40GB limit, and using it for job data storage violates the RC storage policy. Always specify a data directory using --data-dir to avoid quota issues and policy violations.

Label Studio is available on HyPerGator and below you can see walkthroughs on 2 ways to easily access label-studio and connect it to your own data directory.

Method 1: SSH Port Forwarding¶

SSH into HyPerGator with port forwarding:
```
ssh -L <ANY_OPEN_LOCAL_MACHINE_PORT>:localhost:8080 <username>@hpg.rc.ufl.edu
```
> Note: if you don't know what to enter for ANY_OPEN_LOCAL_MACHINE_PORT, you can enter 8081
Load the module and start Label Studio (specifying a data dir to store all label studio data):
```
module load label-studio
label-studio start --data-dir <YOUR_DATA_DIRECTORY>
```
> Note: It is strongly recommended that you specify a large data directory since otherwise Label Studio will default to your home directory.
Open your local browser and navigate to http://localhost:<ANY_OPEN_LOCAL_MACHINE_PORT> which should bring you to your instance of label studio.

Method 2: OnDemand Remote Desktop¶

Go to ood.rc.ufl.edu
Login and create a HiPerGator Desktop session found under interactive apps
Open a terminal inside the desktop environment once available

Load the module and start Label Studio:

module load label-studio
label-studio start --data-dir <YOUR_DATA_DIRECTORY>

Label Studio will open in the HiPerGator Desktop browser

Data Directory Setup¶

Important: Specify a data directory to persist your work and share data with team members anytime you start label studio:

label-studio start --data-dir /path/to/project/data

This ensures your annotations and project settings are saved and accessible for future sessions. It will also prevent accidentally using your home directory for job data storage which violates RC storage policies.

Categories¶

annotation, data_science, image_processing